[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

2.6-test2 on unusual UE2



I have a rather strange 3rd party UE2 I'm trying to get debian and the latest 
linux working on.
Here's the header from the boot prom:
SPARCengine(tm)Ultra(tm) CycleQUAD (2 X UltraSPARC-II 400MHz), No Keyboard
OpenBoot 3.10, 2048 MB memory installed, Serial #31158.

This is a PCI based UE2 clone.  Room for 4 CPUs.

I'm trying to compile kernels with egcs64 (egcs-2.92.11) and gcc 3.0.4 (both 
defaults from debian)
I've managed to get a config that mostly works, but I have two problems left.
The first is a bit of messiness with my drives.
Linux is discovering them in the wrong order.
I have the 2 drives in the machine itself, 8 drives hooked to a differential 
controller and D1000, and 4 more drives hooked to another PCI based SCSI 
controller. 
(scsi4 : QLogic ISP1020 SCSI on PCI bus 01 device 20 irq 7039488 MEM base 0x1ff80)

The problem is that the 2.6 kernel is finding the PCI Qlogic controller first, 
so the boot drive comes up as /dev/sde and that doesn't work out very well. 
(to say the least)
Is there anyway to shift the order of discovery so the internal drives are 
found first?
(the other controllers are Sym, 2 internal, and 2 on the differential card)

The second issue, is of course, SMP.
I can compile an SMP kernel just fine, but it just doesn't boot.
I get:
Remapping the kernel... done.
Booting Linux...

and then silence.
Anyone have any ideas on this one?
Should I be using a different combo of compilers to make it?

Greg




Reply to: